Codified audio language modeling learns useful representations for music information retrieval

Open paper
Benchmark Model Rank Results
emotion-recognition-on-emomusicJukebox (Pre-training: CALM)#4EmoA: 72.1EmoV: 61.7
emotion-recognition-on-emomusicCLMR (Pre-training: contrastive)#5EmoA: 67.8EmoV: 45.8

Rank counts only results with a code link.